Bug #116576 InnoDB Deadlock and Assertion Failure Leading to MySQL Crash
Submitted: 7 Nov 7:58 Modified: 28 Nov 14:39
Reporter: Arpit Arpit Email Updates:
Status: Need Feedback Impact on me:
None 
Category:MySQL Server: Replication Severity:S1 (Critical)
Version:8.0.37 OS:CentOS
Assigned to: MySQL Verification Team CPU Architecture:Any

[7 Nov 7:58] Arpit Arpit
Description:
We are experiencing frequent deadlocks and crashes on MySQL 8.0.37 slave servers in a master-slave setup due to InnoDB assertion failures, resulting in MySQL termination without query logs to indicate the cause; despite enabling the Performance Schema on the slave servers, no insights were gained, and the error logs show mutex contention and fatal errors in sync0arr.cc—this is impacting replication stability and reliability.

Issue Details

Deadlock Occurrences: The deadlocks occur frequently on the slave servers, disrupting the replication process and causing instability.
Lack of Query Logs: During these deadlock events, the MySQL error log does not capture any information about the queries involved in the deadlock, making it challenging to trace the issue.
Performance Schema: After enabling the Performance Schema on the slave servers, we still did not find any useful insights into the deadlock sources.

How to repeat:
The issue occurs intermittently, especially under high load. However, specific steps to reliably reproduce it are unclear, as it does not follow a predictable pattern.
[7 Nov 11:55] MySQL Verification Team
Hi,

Please upload your full config of all nodes, show global status/variables from all nodes and your full log files from all nodes.

Did you change anything before this issue started occurring?

Thanks
[28 Nov 14:39] MySQL Verification Team
Hi,

I have problem reproducing this on 8.0.37.

I see you have some errors about XA transactions, are you using XA transactions too?